The field of the disclosure relates generally to visual analysis of objects, and, more particularly, to tracking and classifying objects of interest in a video.
Existing systems for visual detection of objects are generally based on some sort of distinctive characteristic of an object as compared to a disregarded background. The distinctive characteristics may be color, spatial, or movement characteristics compared to the background. Such systems may be somewhat robust at object detection, but require a human operator to classify (i.e., identify) the object. Further, although at least some known systems are able to classify objects, such systems may only be able to classify an entire image as one object, instead of segmenting the image into discrete objects. Moreover, at least some known systems iteratively search an entire image in smaller sub-sections to classify discrete objects, which requires significant computational time and resources.